An Optimal Balanced Partitioning of a Set of 1D Intervals
|
|
- Oliver Anderson
- 7 years ago
- Views:
Transcription
1 Intentional Journal of Artificial Life Research An Optimal Balanced Partitioning of a Set of 1D Intervals CHUAN-KAI YANG* Department of Information Management National Taiwan University of Science and Technology No. 43, Section 4, Keelung Road, Taipei, 10607, Taiwan Abstract Given a set of 1D intervals and a desired partition number, this paper studies on how to make an optimal partitioning of these intervals, such that the number of intervals between the largest partition and smallest partition is minimal among all possible partitioning schemes. Though seemingly easy at the first glance, this problem has its difficulty due to the fact that an interval ``striding'' multiple partitions should be counted multiple times. Previously we have given an approximated solution to this problem by employing a simulated annealing approach [Yang & Chiueh, 2006], which could give satisfactory results in most cases; however, there is no theoretical guarantee on its optimality. In this paper, we propose a method that could both optimally and deterministically partition a given set of 1D intervals into a given number of partitions. This problem originates from dealing with the partition of a large volume data, where a more balanced partitioning could facilitate efficient out-of-core volume visualization faced in the computer graphics world. We will show that some types of load balancing problem could also be formulated as a balanced interval partitioning problem. As a result, seeking better solutions to this problem has become an important issue. Keywords: Computational Geometry, Dynamic Programming, Interval Partitioning 1. MOTIVATION AND INTRODUCTION How to partition a set of 1D intervals into balanced partitions is what to be discussed in this paper, and such a problem originally arisen from dealing with efficient volume visualization. Volume visualization, sometimes deemed a sub-field of computer graphics, aims to provide 2D images, or more specifically, 2D visualizations, from a so called 3D volume dataset, which is normally composed of a stack of 2D images. Instead of distributing data values on 2D grid points like a 2D image, a volume dataset generalizes the distribution to 3D grid points. Iso-surface extraction is one of the most popular methods to visualize a volume dataset, where an iso-value is first given, and then the corresponding iso-surface, representing the surface intersecting with the given value, is extracted and represented by a set of interpolated triangles. These triangles could later be sent to a graphics card to be rendered into images. As technology advances, nowadays data acquisition devices tend to collect or generate data with higher and higher precision, thus leading to volume datasets of huge sizes. The visualization of a huge volume dataset 1
2 2 CHUAN-KAI YANG usually would cause problems as such a task normally needs to bring in the whole dataset into the main memory before any further processing could start. As the size of a huge volume dataset could be more than several GBytes, the scenario of an out-of-core processing arises [Cox & Ellsworth, 1997; Cox, 1997; Ueng et~al., 1997], which means the required memory for data processing is larger than a system could provide. To address this, data partitioning is one popular approach, such as the method proposed by Chiang et al. [Chiang et~al., 1998]. Since volume visualization is often the most important purpose that a dataset was collected or generated in the first place, it should be taken into account when dealing with data partitioning. In Chiang et al.'s approach, they partition a volume dataset spatially so that each resulting partition could be loaded into the main memory and processed individually without incurring any virtual memory overheads, namely, memory swappings or pagings. The extracted iso-surface is an aggregate result from each partition. Such an paradigm, though simple, has the following drawback: unless some clever pre-processing is employed, all the partitions need to be examined for iso-surface intersection. We argue that an different partitioning scheme may sometimes offer better processing efficiency. That is, instead of partitioning a volume dataset based on spatial information, we could decompose a dataset into layers based on the data values associated with the grid points, and each layer represents a portion of data bearing a particular value range. Such a partitioning has the advantage that for an iso-value query, at most one layer is needed for extracting the iso-surface, thus making it an attractive scheme. Furthermore, the result is a continuous or integrated representation, rather than a segmented one faced in the case of a spatial partitioning, where a post stitching or merging may be needed for smoothing or reducing the rendering traffic sent to graphics cards afterward. However, there are also two issues associated with this kind of partitioning approach. First, it is not suitable for a volume dataset with high data variation, such as some of the datasets obtained from CFD (computational fluid dynamics) simulation, since a value-based partitioning may lead to decompositions that are too fragmented, thus incurring extra representation overheads. Second, how to make a balanced partitioning is in fact more difficult that it seems, as will be detailedly explained in later sections. For the ease of ensuing discussion, we will refer to the first type of partitioning a space-based or coordinate-based scheme, while the second type of partitioning a value-based or layer-based scheme. Previously [Yang & Chiueh, 2006] have shown that how a balanced (with a slightly different definition of balance) partitioning could be achieved by a simulated annealing approach. Though efficient, such a soft optimization method can neither guarantee the optimality nor provide a theoretical analysis of the result. In this work, we provide a dynamic programming approach that could find an optimal solution to this problem. Applications of this balanced interval partitioning problem are not limited to volume visualization only, as some types of load balancing problems could also be formulated as such a problem. For example, considering the following load balancing problem, where a number of activities are to be held in a hall, and each activity has its own specific starting time and duration. Now assuming there are only a fixed number of janitors available to assist these activities, a reasonable arrangement is to partition these activities into groups so that each janitor is responsible for roughly the same number of activities. Some balanced task scheduling problems could also be formulated in a similar fashion, thus making solutions to this problem more important.
3 An Optimal Balanced Partitioning of a Set of 1D Intervals 3 The rest of the paper is organized as the following. Section 2 reviews the literature related to this work. Section 3 details our proposed algorithm. Section 4 analyzes the time complexity of our proposed approach. Section 5 concludes this work and points out some potential future directions. 2. RELATED WORK As explained in the Introduction Section, the problem of balanced 1D interval partitioning comes from the field of volume visualization, and the visualization method involved is called iso-surface extraction, proposed by Lorensen et al. [Lorensen & Cline, 1987]. An iso-surface extraction process identifies the surface from a volume dataset that intersects with a given iso-value. Such an iso-surface query is basically a stabbing query and the original concept came from Edelsbrunner [Edelsbrunner, 1980] and McCreight [McCreight, 1980] in the computational geometry community, and there are ways to speed up this type of queries, such as the proposal of using interval trees [Cignoni et~al., 1997]. However, as a volume dataset could get larger and larger, or sometimes even larger than a machine's physical memory, the so called out-of-core situation would arise, and some dataset partitioning scheme may seem unavoidable. Chiang et al. [Chiang et~al., 1998] partitions a volume dataset into meta-cells using a KD-tree [Bentley, 1975] decomposition to make sure that each meta-cell could be loaded into the memory. Ma et al. proposed a similar scheme [Ma et~al., 2001] by spatially decomposing a volume dataset into layers, which should be noted to bear a different meaning from ours. Yang et al. [Yang & Chiueh, 2006] solves a similar interval partitioning problem by a simulated-annealing approach. By treating the number of intervals in a partition as a random variable, their approach to achieve a balanced partitioning is to minimize the variance of such random variables. Note that, with a straightforward modification, the minimization of variance could be easily converted to the minimization of the size difference between the largest and the smallest partitions. 3. A BALANCED LAYER PARTITIONING ALGORITHM We formally define what our target problem is and propose our solution to it in this Section. 3.1 Problem Statement Given a set of arbitrary 1D intervals, represented by I, and a partition number, represented by M, we want to partition I into M partitions such that in terms of number of intervals, the size difference between the largest and smallest partitions, is minimal. 3.2 A Divide-and-Conquer Approach To illustrate why such a problem is more difficult than it seems, we first give a naive divide-and-conquer approach. Assume that 2, apparently the problem becomes to find one cut to make two partitions such that the difference between the numbers of intervals intersecting with each partition is minimal. Note that as some intervals
4 4 CHUAN-KAI YANG may intersect with the cut itself, these intersecting intervals will be counted twice, i.e., they will contribute to both partitions. In other words, making a cut will ``interfere'' with the number of intervals intersecting with the affected partitions, and this is exactly why this problem is intricate. To proceed, we could adopt the following procedure which is intrinsically a binary search algorithm. The lower bound and upper bound of all intervals are first determined, denoted by Min and Max, respectively, and then we calculate the middle point, or the initial position for the cut, denoted by Mid, where. Let the number of intervals intersecting with the closed interval [Min, Mid] be Left and similarly the number of intervals intersecting with the closed interval [Mid, Max] be Right, where [a, b] denotes the set of real numbers:. If, then we are done. On the other hand if, the new cutting position is set to. Finally for the remaining case, the new cutting position is set to. Such a procedure could be carried out until the difference between Left and Right could not be further reduced. Now let us turn to the case where 4. A naive or straightforward approach would be to first make two partitions, denoted by and, respectively, as in the case where 2. Then we further partition into and by making one cut, denoted by. Similarly denotes the cut to partition into and. Now the problem arises. As the number of intervals intersecting with and may be quite different, and the former will contribute to and, while the latter to and, thus leading to an overall imbalanced partitioning. Therefore a divide-and-conquer approach like this would fail. However, such an approach may be used as a starting configuration for further adjustment to achieve a balanced partitioning, such as the simulated annealing approach proposed by Yang et al. [Yang & Chiueh, 2006]. 3.3 A Dynamic Programming Approach To solve this problem deterministically, in this paper, we resort to a dynamic programming approach, which is very similar to the one used to solve the matrix chain multiplication problem [Cormen et~al., 2001]. The recursive definition of the dynamic programming is shown in Equation~1 while the C-like pseudo code is listed in Figure 1, where some terms should be explained in advance. First, M denotes the desired number of partitions, and Num_I is the number of intervals, while N represents the number of distinct 1D points coming from the corresponding end points of these Num_I intervals. To make M partitions, all we need to do is to determine 1 cutting positions. However, it does not take long for us to figure out that there are in fact at most 2 1 positions need to be considered. That is, the positions of the N points, together with the 1 midpoints between any two adjacent points. The configuration produced by making any other cutting position could be shown to be equivalent to placing a cut at one of the above 2 1 positions. For the ease of ensuing discussion, we sort these 2 1 positions from left to right, and will number them from 1 to 2 1. Second, the term,, stands for the best solution of making k partitions of the region starting from the i-th position to the j-th position. Therefore,,. _ represents the cutting position to further decompose this region into two partitions, say and, and,,. _ is the number of partitions to be made in
5 An Optimal Balanced Partitioning of a Set of 1D Intervals 5, thus leaving,,. _ partitions to be made in. Furthermore, as the names suggest,,,. and,,. represent the numbers of intervals in the minimal and maximal partitions, respectively, and by definition, the difference between these two numbers is minimal among all possible partitioning schemes to make k partition in this region. Finally, MIN and MAX are two macros to extract the smaller and larger number, respectively, when two numbers are given. As mentioned previously, the basic flow of this dynamic programming approach follows the matrix chain multiplication problem closely by building up the calculations of,, in a bottom-up manner. 0 0,,, 1 1,,,,, where,,,, is obtained by calculating,,.,,,.,,.,,,. 2 for(k=2;k<=m;k++) { for(len=1;len<=2*n-1;len++) { for(i=1;i<2*n-l;i++) { j=i+len; diff_temp=num_i*m; for(k1=1;k1<k;k1++) { for(i1=i;i1<=j;i1++) { min_temp=min(part(i,i1,k-k1).min,part(i1,j,k1).min); max_temp=max(part(i,i1,k-k1).max,part(i1,j,k1).max); if(max_temp-min_temp<diff_temp) { diff_temp=max_temp-min_temp; Part(i,j,k).part_pos=i1; Part(i,j,k).part_num=k-k1; Part(i,j,k).min=min_temp; Part(i,j,k).max=max_temp; Fig. 1. The pseudo code of a dynamic programming approach. The code in Figure 1 presents a five-loop structure, where the first loop increases the partition number, i.e., the term k in,,, while the second loop controls the length, represented as len in the code, of each involved partition, i.e., how far apart i and j are in,,. Note that in our code, j is never less than i, and thus is always set to
6 6 CHUAN-KAI YANG be. The third loop enumerates all possible positions of i in,,, together with the previously enumerated length, therefore the region to be partitioned in,, is completely determined. The fourth and fifth loops are used to divide the partition of,, into two smaller partitions where the left one, denoted by,,, and the right one, denoted by,,, are compared to find the best values of and such that the combined results from both partitions could form the most balanced partitioning, as far as,, is concerned. The search for optimality is to minimize the difference, denoted by diff_temp, between the largest partition and the smallest partition, and is initialized to be Num_I M, which is apparently an upper bound that could not be exceeded. The initialization of the pseudo code is shown in Figure 2, where the bottommost values of,, are laid down, and such an initial setting should be self-explanatory, except the term,. Here, is used to record the number of intervals intersecting with the region defined by the i-th and j-th cutting positions, under the constraint that. for(i=1;i<2*n;i++) { for(j=i;j<2*n;j++) { Part(i,j,1).min=part(i,j,1).max=query(i,j); Part(i,j,1).part_num=0; Part(i,j,1).part_pos=i; for(i=1;i<2*n;i++) for(k=2;k<=m;k++) { Part(i,i,k).min=part(i,i,k).max=query(i,i); Part(i,i,k).part_num=0; Part(i,i,k).part_pos=i; Fig. 2. The initialization of the pseudo code in Figure 1. Note that, compared with the naive divide-and-conquer approach, where a cut could introduce side-effects and therefore may complicate the partitioning, this dynamic programming algorithm takes into account the cutting overheads by counting values in the form of,,, through which the numbers of intervals intersecting with the cuts will be handled implicitly and correctly. 4. COMPLEXITY ANALYSIS In this section, we analyze the time complexity of our proposed algorithm, which is composed of two parts, as shown in Figure 1 and Figure 2. It is evident that the first part demands an time complexity. On the other hand, the time complexity of the second part depends on how we calculate, each time, given different combinations of i and j. To have a more efficient implementation, we choose to calculate and store all possible, in advance, therefore in Figure 2 the query result of, could be answered with a constant time complexity. As for the total time complexity involved to pre-calculate all possible values of,, it could be
7 An Optimal Balanced Partitioning of a Set of 1D Intervals 7 shown by the following argument that it is bounded by. To begin with, we first define as the number of intervals use the i-th position as their left end point. All values could be calculated in time as we could just enumerate each interval and collect required information accordingly. Secondly, it is not difficult to see why the equality,, 1 holds, as the inclusion of the i-th position brings in exactly new intervals, with respect to the, 1, intervals that are already counted. This indicates that the calculation of, could be carried out by an iterative computation. However, such an iteration requires the value of,, which is exactly the form of a stabbing query and could be answered in time by building an interval tree at the pre-processing time with an total time complexity [Cignoni et~al., 1997]. As the forms of, are only calculated times, the total time complexity in Figure 2 is therefore. Consequently, the overall algorithm complexity is dominated by the first part, and thus is. Note that the associated space complexity, evidently dominated by the memorization of,,, is. 5 CONCLUSIONS AND FUTURE WORK We have proposed a deterministic algorithm to find an optimal solution to the interval partitioning problem. There are two future directions that we plan to pursue. First, how to determine the most suitable number of partitions for a given dataset is worth studying, as a partitioning cut could cause extra intersection overhead, and therefore too many partitions may lead to an overall inefficient scheme for data storage. Potential solutions may include some hierarchical data structures or hybrid schemes of space-based and value-based approaches. Second, we must admit that an overall time complexity is still too high, especially when we are faced with huge datasets where the associated values of N larger than one million are not uncommon. There exist techniques to reduce the time complexity of the matrix chain multiplication problem from [Cormen et~al., 2001] to [Hu & Shing, 1982; Hu & Shing, 1984] by employing a divide-and-conquer approach, and due to its similarity to our problem, we are currently surveying the possibility for time complexity reduction in this regard. Furthermore, space complexity reduction is also a very important concern, as a huge dataset may render the whole algorithm inoperable, as the required memory cannot be provided. REFERENCES [Bentley, 1975] Bentley, J.~L. (1975). Multidimensional Binary Search Trees Used for Associative Searching. Communications of the ACM, 18(9): [Chiang et~al., 1998] Chiang, Y., Silva, C., and Schroeder, W. (1998). Interactive Out-of-Core Isosurface Extraction. In IEEE Visualization '98, pages [Cignoni et~al., 1997] Cignoni, P., Marino, P., Montani, C., Puppo, E., and Scopigno, R. (1997). Speeding Up Isosurface Extraction Using Interval Trees. IEEE Transactions on Visualization and Computer Graphics, 3(2):
8 8 CHUAN-KAI YANG [Cormen et~al., 2001] Cormen, T.~H., Leiserson, C.~E., and Rivest, R.~L. (2001). Introduction to Algorithms. MIT Press, second edition. [Cox, 1997] Cox, M. (1997). Managing Big Data for Scientific Visualization. ACM SIGGRAPH '98 Course. [Cox and Ellsworth, 1997] Cox, M. and Ellsworth, D. (1997). Application-Controlled Demand Paging for Out-of-core Visualization. IEEE Visualization '97, pages [Edelsbrunner, 1980] Edelsbrunner, H. (1980). Dynamic Data Structures for Orthogonal Intersection Queries. Technical Report Report F59, Inst. Informationsverarb.,Tech. University Graz. [Hu and Shing, 1982] Hu, T.~C. and Shing, M.~T. (1982). Computation of matrix chain products. part i. SIAM Journal on Computing, 11(2): [Hu and Shing, 1984] Hu, T.~C. and Shing, M.~T. (1984). Computation of matrix chain products. part ii. SIAM Journal on Computing, 13(2): [Lorensen and Cline, 1987] Lorensen, W.~E. and Cline, H.~E. (1987). Marching Cube: A High Resolution 3D Surface Construction Algorithm. Computer Graphics, 21(4): [Ma et~al., 2001] Ma, K., Abla, G., and Lum, E. (2001). Layer Data Organization for Visualizing Unstructured-Grid Data. In Proceedings of SPIE, Visual Data Exploration and Analysis VIII, pages [McCreight, 1980] McCreight, E.~M. (1980). Efficient Algorithms for Enumerating Intersecting Intervals and Rectangles. Technical Report Report CSL-80-9, Xerox Palo Alto Res. Center. [Ueng et~al., 1997] Ueng, S.~K., Siborski, K., and Ma, K.~L. (1997). Out-of-Core Streamline Visualization on Large unstructured meshes. ICASE Report. [Yang and Chiueh, 2006] Yang, C. and Chiueh, T. (2006). Integration of Volume Decompression and Out-of-Core Iso-Surface Extraction from Irregular Volume Data. The Visual Computer, 22(4):
Junghyun Ahn Changho Sung Tag Gon Kim. Korea Advanced Institute of Science and Technology (KAIST) 373-1 Kuseong-dong, Yuseong-gu Daejoen, Korea
Proceedings of the 211 Winter Simulation Conference S. Jain, R. R. Creasey, J. Himmelspach, K. P. White, and M. Fu, eds. A BINARY PARTITION-BASED MATCHING ALGORITHM FOR DATA DISTRIBUTION MANAGEMENT Junghyun
More informationOut-Of-Core Algorithms for Scientific Visualization and Computer Graphics
Out-Of-Core Algorithms for Scientific Visualization and Computer Graphics Course Notes for IEEE Visualization 2002 Boston, Massachusetts October 2002 Organizer: Cláudio T. Silva Oregon Health & Science
More informationCompact Representations and Approximations for Compuation in Games
Compact Representations and Approximations for Compuation in Games Kevin Swersky April 23, 2008 Abstract Compact representations have recently been developed as a way of both encoding the strategic interactions
More informationA Generalized Marching Cubes Algorithm Based On Non-Binary Classifications
Konrad-Zuse-Zentrum fu r Informationstechnik Berlin Takustraße 7 D-14195 Berlin-Dahlem Germany HANS-CHRISTIAN HEGE DETLEV STALLING MARTIN SEEBASS MALTE ZOCKLER A Generalized Marching Cubes Algorithm Based
More informationIMPROVING PERFORMANCE OF RANDOMIZED SIGNATURE SORT USING HASHING AND BITWISE OPERATORS
Volume 2, No. 3, March 2011 Journal of Global Research in Computer Science RESEARCH PAPER Available Online at www.jgrcs.info IMPROVING PERFORMANCE OF RANDOMIZED SIGNATURE SORT USING HASHING AND BITWISE
More informationVolume visualization I Elvins
Volume visualization I Elvins 1 surface fitting algorithms marching cubes dividing cubes direct volume rendering algorithms ray casting, integration methods voxel projection, projected tetrahedra, splatting
More informationFast Sequential Summation Algorithms Using Augmented Data Structures
Fast Sequential Summation Algorithms Using Augmented Data Structures Vadim Stadnik vadim.stadnik@gmail.com Abstract This paper provides an introduction to the design of augmented data structures that offer
More informationSECONDARY STORAGE TERRAIN VISUALIZATION IN A CLIENT-SERVER ENVIRONMENT: A SURVEY
SECONDARY STORAGE TERRAIN VISUALIZATION IN A CLIENT-SERVER ENVIRONMENT: A SURVEY Kai Xu and Xiaofang Zhou School of Information Technology and Electrical Engineering The University of Queensland, Brisbane,
More informationLecture 7 - Meshing. Applied Computational Fluid Dynamics
Lecture 7 - Meshing Applied Computational Fluid Dynamics Instructor: André Bakker http://www.bakker.org André Bakker (2002-2006) Fluent Inc. (2002) 1 Outline Why is a grid needed? Element types. Grid types.
More informationConstrained Tetrahedral Mesh Generation of Human Organs on Segmented Volume *
Constrained Tetrahedral Mesh Generation of Human Organs on Segmented Volume * Xiaosong Yang 1, Pheng Ann Heng 2, Zesheng Tang 3 1 Department of Computer Science and Technology, Tsinghua University, Beijing
More informationSolving Geometric Problems with the Rotating Calipers *
Solving Geometric Problems with the Rotating Calipers * Godfried Toussaint School of Computer Science McGill University Montreal, Quebec, Canada ABSTRACT Shamos [1] recently showed that the diameter of
More informationMapReduce Approach to Collective Classification for Networks
MapReduce Approach to Collective Classification for Networks Wojciech Indyk 1, Tomasz Kajdanowicz 1, Przemyslaw Kazienko 1, and Slawomir Plamowski 1 Wroclaw University of Technology, Wroclaw, Poland Faculty
More informationCategorical Data Visualization and Clustering Using Subjective Factors
Categorical Data Visualization and Clustering Using Subjective Factors Chia-Hui Chang and Zhi-Kai Ding Department of Computer Science and Information Engineering, National Central University, Chung-Li,
More informationBernice E. Rogowitz and Holly E. Rushmeier IBM TJ Watson Research Center, P.O. Box 704, Yorktown Heights, NY USA
Are Image Quality Metrics Adequate to Evaluate the Quality of Geometric Objects? Bernice E. Rogowitz and Holly E. Rushmeier IBM TJ Watson Research Center, P.O. Box 704, Yorktown Heights, NY USA ABSTRACT
More informationSolving Linear Programs
Solving Linear Programs 2 In this chapter, we present a systematic procedure for solving linear programs. This procedure, called the simplex method, proceeds by moving from one feasible solution to another,
More informationRegular Expressions and Automata using Haskell
Regular Expressions and Automata using Haskell Simon Thompson Computing Laboratory University of Kent at Canterbury January 2000 Contents 1 Introduction 2 2 Regular Expressions 2 3 Matching regular expressions
More informationTime series clustering and the analysis of film style
Time series clustering and the analysis of film style Nick Redfern Introduction Time series clustering provides a simple solution to the problem of searching a database containing time series data such
More informationSeveral Views of Support Vector Machines
Several Views of Support Vector Machines Ryan M. Rifkin Honda Research Institute USA, Inc. Human Intention Understanding Group 2007 Tikhonov Regularization We are considering algorithms of the form min
More informationCHAPTER 1. Introduction to CAD/CAM/CAE Systems
CHAPTER 1 1.1 OVERVIEW Introduction to CAD/CAM/CAE Systems Today s industries cannot survive worldwide competition unless they introduce new products with better quality (quality, Q), at lower cost (cost,
More informationLoad Balancing on a Grid Using Data Characteristics
Load Balancing on a Grid Using Data Characteristics Jonathan White and Dale R. Thompson Computer Science and Computer Engineering Department University of Arkansas Fayetteville, AR 72701, USA {jlw09, drt}@uark.edu
More informationOffline sorting buffers on Line
Offline sorting buffers on Line Rohit Khandekar 1 and Vinayaka Pandit 2 1 University of Waterloo, ON, Canada. email: rkhandekar@gmail.com 2 IBM India Research Lab, New Delhi. email: pvinayak@in.ibm.com
More informationComputational Geometry. Lecture 1: Introduction and Convex Hulls
Lecture 1: Introduction and convex hulls 1 Geometry: points, lines,... Plane (two-dimensional), R 2 Space (three-dimensional), R 3 Space (higher-dimensional), R d A point in the plane, 3-dimensional space,
More informationMultiresolution Terrain Database Visualization
Multiresolution Terrain Database Visualization Kai Xu School of Information Technology and Electrical Engineering The University of Queensland, Brisbane, QLD 4072, Australia kaixu@itee.uq.edu.au Abstract.
More informationVolumetric Meshes for Real Time Medical Simulations
Volumetric Meshes for Real Time Medical Simulations Matthias Mueller and Matthias Teschner Computer Graphics Laboratory ETH Zurich, Switzerland muellerm@inf.ethz.ch, http://graphics.ethz.ch/ Abstract.
More information1 Review of Least Squares Solutions to Overdetermined Systems
cs4: introduction to numerical analysis /9/0 Lecture 7: Rectangular Systems and Numerical Integration Instructor: Professor Amos Ron Scribes: Mark Cowlishaw, Nathanael Fillmore Review of Least Squares
More informationOff-line Model Simplification for Interactive Rigid Body Dynamics Simulations Satyandra K. Gupta University of Maryland, College Park
NSF GRANT # 0727380 NSF PROGRAM NAME: Engineering Design Off-line Model Simplification for Interactive Rigid Body Dynamics Simulations Satyandra K. Gupta University of Maryland, College Park Atul Thakur
More informationScheduling Shop Scheduling. Tim Nieberg
Scheduling Shop Scheduling Tim Nieberg Shop models: General Introduction Remark: Consider non preemptive problems with regular objectives Notation Shop Problems: m machines, n jobs 1,..., n operations
More informationBachelor of Games and Virtual Worlds (Programming) Subject and Course Summaries
First Semester Development 1A On completion of this subject students will be able to apply basic programming and problem solving skills in a 3 rd generation object-oriented programming language (such as
More informationClosest Pair Problem
Closest Pair Problem Given n points in d-dimensions, find two whose mutual distance is smallest. Fundamental problem in many applications as well as a key step in many algorithms. p q A naive algorithm
More informationOnline Scheduling for Cloud Computing and Different Service Levels
2012 IEEE 201226th IEEE International 26th International Parallel Parallel and Distributed and Distributed Processing Processing Symposium Symposium Workshops Workshops & PhD Forum Online Scheduling for
More informationBuffer Operations in GIS
Buffer Operations in GIS Nagapramod Mandagere, Graduate Student, University of Minnesota npramod@cs.umn.edu SYNONYMS GIS Buffers, Buffering Operations DEFINITION A buffer is a region of memory used to
More informationThe Design and Implement of Ultra-scale Data Parallel. In-situ Visualization System
The Design and Implement of Ultra-scale Data Parallel In-situ Visualization System Liu Ning liuning01@ict.ac.cn Gao Guoxian gaoguoxian@ict.ac.cn Zhang Yingping zhangyingping@ict.ac.cn Zhu Dengming mdzhu@ict.ac.cn
More informationWavelet Representation of functions defined on Tetrahedrical Grids
Wavelet Representation of functions defined on Tetrahedrical Grids Silvia Castro (1), Liliana Castro (2), Liliana Boscardín (2), and Armando De Giusti (3) (1) Dpto. de Cs. e Ing. de la Computación (2)
More informationPartitioning and Divide and Conquer Strategies
and Divide and Conquer Strategies Lecture 4 and Strategies Strategies Data partitioning aka domain decomposition Functional decomposition Lecture 4 and Strategies Quiz 4.1 For nuclear reactor simulation,
More informationMemory Allocation Technique for Segregated Free List Based on Genetic Algorithm
Journal of Al-Nahrain University Vol.15 (2), June, 2012, pp.161-168 Science Memory Allocation Technique for Segregated Free List Based on Genetic Algorithm Manal F. Younis Computer Department, College
More informationBEHAVIOR BASED CREDIT CARD FRAUD DETECTION USING SUPPORT VECTOR MACHINES
BEHAVIOR BASED CREDIT CARD FRAUD DETECTION USING SUPPORT VECTOR MACHINES 123 CHAPTER 7 BEHAVIOR BASED CREDIT CARD FRAUD DETECTION USING SUPPORT VECTOR MACHINES 7.1 Introduction Even though using SVM presents
More informationCHAPTER 1 INTRODUCTION
CHAPTER 1 INTRODUCTION Power systems form the largest man made complex system. It basically consists of generating sources, transmission network and distribution centers. Secure and economic operation
More informationL25: Ensemble learning
L25: Ensemble learning Introduction Methods for constructing ensembles Combination strategies Stacked generalization Mixtures of experts Bagging Boosting CSCE 666 Pattern Analysis Ricardo Gutierrez-Osuna
More informationA HYBRID GROUND DATA MODEL TO SUPPORT INTERACTION IN MECHANIZED TUNNELING
A HYBRID GROUND DATA MODEL TO SUPPORT INTERACTION IN MECHANIZED TUNNELING F. HEGEMANN P. MANICKAM K. LEHNER M. KÖNIG D. HARTMANN Department of Civil and Environmental Engineering, Ruhr-University of Bochum,44780
More informationA Note on Maximum Independent Sets in Rectangle Intersection Graphs
A Note on Maximum Independent Sets in Rectangle Intersection Graphs Timothy M. Chan School of Computer Science University of Waterloo Waterloo, Ontario N2L 3G1, Canada tmchan@uwaterloo.ca September 12,
More informationA Strategy for Teaching Finite Element Analysis to Undergraduate Students
A Strategy for Teaching Finite Element Analysis to Undergraduate Students Gordon Smyrell, School of Computing and Mathematics, University of Teesside The analytical power and design flexibility offered
More informationHow To Find Influence Between Two Concepts In A Network
2014 UKSim-AMSS 16th International Conference on Computer Modelling and Simulation Influence Discovery in Semantic Networks: An Initial Approach Marcello Trovati and Ovidiu Bagdasar School of Computing
More informationAn Early Attempt at Applying Deep Reinforcement Learning to the Game 2048
An Early Attempt at Applying Deep Reinforcement Learning to the Game 2048 Hong Gui, Tinghan Wei, Ching-Bo Huang, I-Chen Wu 1 1 Department of Computer Science, National Chiao Tung University, Hsinchu, Taiwan
More informationANIMATION a system for animation scene and contents creation, retrieval and display
ANIMATION a system for animation scene and contents creation, retrieval and display Peter L. Stanchev Kettering University ABSTRACT There is an increasing interest in the computer animation. The most of
More informationRegions in a circle. 7 points 57 regions
Regions in a circle 1 point 1 region points regions 3 points 4 regions 4 points 8 regions 5 points 16 regions The question is, what is the next picture? How many regions will 6 points give? There's an
More informationMethodology for Emulating Self Organizing Maps for Visualization of Large Datasets
Methodology for Emulating Self Organizing Maps for Visualization of Large Datasets Macario O. Cordel II and Arnulfo P. Azcarraga College of Computer Studies *Corresponding Author: macario.cordel@dlsu.edu.ph
More informationClustering. Danilo Croce Web Mining & Retrieval a.a. 2015/201 16/03/2016
Clustering Danilo Croce Web Mining & Retrieval a.a. 2015/201 16/03/2016 1 Supervised learning vs. unsupervised learning Supervised learning: discover patterns in the data that relate data attributes with
More informationLecture 3: Finding integer solutions to systems of linear equations
Lecture 3: Finding integer solutions to systems of linear equations Algorithmic Number Theory (Fall 2014) Rutgers University Swastik Kopparty Scribe: Abhishek Bhrushundi 1 Overview The goal of this lecture
More informationIntroduction to Algorithms March 10, 2004 Massachusetts Institute of Technology Professors Erik Demaine and Shafi Goldwasser Quiz 1.
Introduction to Algorithms March 10, 2004 Massachusetts Institute of Technology 6.046J/18.410J Professors Erik Demaine and Shafi Goldwasser Quiz 1 Quiz 1 Do not open this quiz booklet until you are directed
More informationTriangulation by Ear Clipping
Triangulation by Ear Clipping David Eberly Geometric Tools, LLC http://www.geometrictools.com/ Copyright c 1998-2016. All Rights Reserved. Created: November 18, 2002 Last Modified: August 16, 2015 Contents
More informationReflection and Refraction
Equipment Reflection and Refraction Acrylic block set, plane-concave-convex universal mirror, cork board, cork board stand, pins, flashlight, protractor, ruler, mirror worksheet, rectangular block worksheet,
More informationOPTIMIZATION OF VENTILATION SYSTEMS IN OFFICE ENVIRONMENT, PART II: RESULTS AND DISCUSSIONS
OPTIMIZATION OF VENTILATION SYSTEMS IN OFFICE ENVIRONMENT, PART II: RESULTS AND DISCUSSIONS Liang Zhou, and Fariborz Haghighat Department of Building, Civil and Environmental Engineering Concordia University,
More informationPredict Influencers in the Social Network
Predict Influencers in the Social Network Ruishan Liu, Yang Zhao and Liuyu Zhou Email: rliu2, yzhao2, lyzhou@stanford.edu Department of Electrical Engineering, Stanford University Abstract Given two persons
More informationNovel Data Extraction Language for Structured Log Analysis
Novel Data Extraction Language for Structured Log Analysis P.W.D.C. Jayathilake 99X Technology, Sri Lanka. ABSTRACT This paper presents the implementation of a new log data extraction language. Theoretical
More informationIntegrated Backup Topology Control and Routing of Obscured Traffic in Hybrid RF/FSO Networks*
Integrated Backup Topology Control and Routing of Obscured Traffic in Hybrid RF/FSO Networks* Abhishek Kashyap, Anuj Rawat and Mark Shayman Department of Electrical and Computer Engineering, University
More informationClustering UE 141 Spring 2013
Clustering UE 141 Spring 013 Jing Gao SUNY Buffalo 1 Definition of Clustering Finding groups of obects such that the obects in a group will be similar (or related) to one another and different from (or
More informationSocial Media Mining. Data Mining Essentials
Introduction Data production rate has been increased dramatically (Big Data) and we are able store much more data than before E.g., purchase data, social media data, mobile phone data Businesses and customers
More informationTHE CFD SIMULATION OF THE FLOW AROUND THE AIRCRAFT USING OPENFOAM AND ANSA
THE CFD SIMULATION OF THE FLOW AROUND THE AIRCRAFT USING OPENFOAM AND ANSA Adam Kosík Evektor s.r.o., Czech Republic KEYWORDS CFD simulation, mesh generation, OpenFOAM, ANSA ABSTRACT In this paper we describe
More informationThe Scientific Data Mining Process
Chapter 4 The Scientific Data Mining Process When I use a word, Humpty Dumpty said, in rather a scornful tone, it means just what I choose it to mean neither more nor less. Lewis Carroll [87, p. 214] In
More informationACL Based Dynamic Network Reachability in Cross Domain
South Asian Journal of Engineering and Technology Vol.2, No.15 (2016) 68 72 ISSN No: 2454-9614 ACL Based Dynamic Network Reachability in Cross Domain P. Nandhini a, K. Sankar a* a) Department Of Computer
More informationBinary Space Partitions
Title: Binary Space Partitions Name: Adrian Dumitrescu 1, Csaba D. Tóth 2,3 Affil./Addr. 1: Computer Science, Univ. of Wisconsin Milwaukee, Milwaukee, WI, USA Affil./Addr. 2: Mathematics, California State
More informationLecture 2: Universality
CS 710: Complexity Theory 1/21/2010 Lecture 2: Universality Instructor: Dieter van Melkebeek Scribe: Tyson Williams In this lecture, we introduce the notion of a universal machine, develop efficient universal
More informationFile System Encryption in C#
INTEGRATED FILE-LEVEL CRYPTOGRAPHICAL ACCESS CONTROL Abstract Ryan Seifert ryanseifert@earthlink.net T. Andrew Yang Yang@UHCL.edu Division of Computing and Mathematics University of Houston - Clear Lake,
More informationFirewall Policy Anomalies- Detection and Resolution
Firewall Policy Anomalies- Detection and Resolution Jitha C K #1, Sreekesh Namboodiri *2 #1 MTech student(cse),mes College of Engineering,Kuttippuram,India #2 Assistant Professor(CSE),MES College of Engineering,Kuttippuram,India
More informationSELECTING NEURAL NETWORK ARCHITECTURE FOR INVESTMENT PROFITABILITY PREDICTIONS
UDC: 004.8 Original scientific paper SELECTING NEURAL NETWORK ARCHITECTURE FOR INVESTMENT PROFITABILITY PREDICTIONS Tonimir Kišasondi, Alen Lovren i University of Zagreb, Faculty of Organization and Informatics,
More informationPredict the Popularity of YouTube Videos Using Early View Data
000 001 002 003 004 005 006 007 008 009 010 011 012 013 014 015 016 017 018 019 020 021 022 023 024 025 026 027 028 029 030 031 032 033 034 035 036 037 038 039 040 041 042 043 044 045 046 047 048 049 050
More informationFAC 2.1: Data Center Cooling Simulation. Jacob Harris, Application Engineer Yue Ma, Senior Product Manager
FAC 2.1: Data Center Cooling Simulation Jacob Harris, Application Engineer Yue Ma, Senior Product Manager FAC 2.1: Data Center Cooling Simulation Computational Fluid Dynamics (CFD) can be used to numerically
More informationGRID SEARCHING Novel way of Searching 2D Array
GRID SEARCHING Novel way of Searching 2D Array Rehan Guha Institute of Engineering & Management Kolkata, India Abstract: Linear/Sequential searching is the basic search algorithm used in data structures.
More informationCall Admission Control and Routing in Integrated Service Networks Using Reinforcement Learning
Call Admission Control and Routing in Integrated Service Networks Using Reinforcement Learning Peter Marbach LIDS MIT, Room 5-07 Cambridge, MA, 09 email: marbach@mit.edu Oliver Mihatsch Siemens AG Corporate
More informationABSTRACT FOR THE 1ST INTERNATIONAL WORKSHOP ON HIGH ORDER CFD METHODS
1 ABSTRACT FOR THE 1ST INTERNATIONAL WORKSHOP ON HIGH ORDER CFD METHODS Sreenivas Varadan a, Kentaro Hara b, Eric Johnsen a, Bram Van Leer b a. Department of Mechanical Engineering, University of Michigan,
More informationModel Repair. Leif Kobbelt RWTH Aachen University )NPUT $ATA 2EMOVAL OF TOPOLOGICAL AND GEOMETRICAL ERRORS !NALYSIS OF SURFACE QUALITY
)NPUT $ATA 2ANGE 3CAN #!$ 4OMOGRAPHY 2EMOVAL OF TOPOLOGICAL AND GEOMETRICAL ERRORS!NALYSIS OF SURFACE QUALITY 3URFACE SMOOTHING FOR NOISE REMOVAL 0ARAMETERIZATION 3IMPLIFICATION FOR COMPLEXITY REDUCTION
More informationNeovision2 Performance Evaluation Protocol
Neovision2 Performance Evaluation Protocol Version 3.0 4/16/2012 Public Release Prepared by Rajmadhan Ekambaram rajmadhan@mail.usf.edu Dmitry Goldgof, Ph.D. goldgof@cse.usf.edu Rangachar Kasturi, Ph.D.
More informationCentralized Systems. A Centralized Computer System. Chapter 18: Database System Architectures
Chapter 18: Database System Architectures Centralized Systems! Centralized Systems! Client--Server Systems! Parallel Systems! Distributed Systems! Network Types! Run on a single computer system and do
More informationAn Overview of Knowledge Discovery Database and Data mining Techniques
An Overview of Knowledge Discovery Database and Data mining Techniques Priyadharsini.C 1, Dr. Antony Selvadoss Thanamani 2 M.Phil, Department of Computer Science, NGM College, Pollachi, Coimbatore, Tamilnadu,
More informationEM Clustering Approach for Multi-Dimensional Analysis of Big Data Set
EM Clustering Approach for Multi-Dimensional Analysis of Big Data Set Amhmed A. Bhih School of Electrical and Electronic Engineering Princy Johnson School of Electrical and Electronic Engineering Martin
More informationThe Combination Forecasting Model of Auto Sales Based on Seasonal Index and RBF Neural Network
, pp.67-76 http://dx.doi.org/10.14257/ijdta.2016.9.1.06 The Combination Forecasting Model of Auto Sales Based on Seasonal Index and RBF Neural Network Lihua Yang and Baolin Li* School of Economics and
More informationFNT EXPERT PAPER. // Data Center Efficiency AUTHOR. Using CFD to Optimize Cooling in Design and Operation. www.fntsoftware.com
FNT EXPERT PAPER AUTHOR Oliver Lindner Head of Business Line DCIM FNT GmbH // Data Center Efficiency Using CFD to Optimize Cooling in Design and Operation Energy is the biggest cost factor with the highest
More informationHSI BASED COLOUR IMAGE EQUALIZATION USING ITERATIVE n th ROOT AND n th POWER
HSI BASED COLOUR IMAGE EQUALIZATION USING ITERATIVE n th ROOT AND n th POWER Gholamreza Anbarjafari icv Group, IMS Lab, Institute of Technology, University of Tartu, Tartu 50411, Estonia sjafari@ut.ee
More informationCSE Case Study: Optimising the CFD code DG-DES
CSE Case Study: Optimising the CFD code DG-DES CSE Team NAG Ltd., support@hector.ac.uk Naveed Durrani University of Sheffield June 2008 Introduction One of the activities of the NAG CSE (Computational
More informationSpatio-Temporal Mapping -A Technique for Overview Visualization of Time-Series Datasets-
Progress in NUCLEAR SCIENCE and TECHNOLOGY, Vol. 2, pp.603-608 (2011) ARTICLE Spatio-Temporal Mapping -A Technique for Overview Visualization of Time-Series Datasets- Hiroko Nakamura MIYAMURA 1,*, Sachiko
More informationLearning in Abstract Memory Schemes for Dynamic Optimization
Fourth International Conference on Natural Computation Learning in Abstract Memory Schemes for Dynamic Optimization Hendrik Richter HTWK Leipzig, Fachbereich Elektrotechnik und Informationstechnik, Institut
More informationDYNAMIC DOMAIN CLASSIFICATION FOR FRACTAL IMAGE COMPRESSION
DYNAMIC DOMAIN CLASSIFICATION FOR FRACTAL IMAGE COMPRESSION K. Revathy 1 & M. Jayamohan 2 Department of Computer Science, University of Kerala, Thiruvananthapuram, Kerala, India 1 revathysrp@gmail.com
More informationMultiresolution 3D Rendering on Mobile Devices
Multiresolution 3D Rendering on Mobile Devices Javier Lluch, Rafa Gaitán, Miguel Escrivá, and Emilio Camahort Computer Graphics Section Departament of Computer Science Polytechnic University of Valencia
More informationLinear Programming. March 14, 2014
Linear Programming March 1, 01 Parts of this introduction to linear programming were adapted from Chapter 9 of Introduction to Algorithms, Second Edition, by Cormen, Leiserson, Rivest and Stein [1]. 1
More informationOperating Systems. Virtual Memory
Operating Systems Virtual Memory Virtual Memory Topics. Memory Hierarchy. Why Virtual Memory. Virtual Memory Issues. Virtual Memory Solutions. Locality of Reference. Virtual Memory with Segmentation. Page
More informationA Hybrid Load Balancing Policy underlying Cloud Computing Environment
A Hybrid Load Balancing Policy underlying Cloud Computing Environment S.C. WANG, S.C. TSENG, S.S. WANG*, K.Q. YAN* Chaoyang University of Technology 168, Jifeng E. Rd., Wufeng District, Taichung 41349
More informationShortest Inspection-Path. Queries in Simple Polygons
Shortest Inspection-Path Queries in Simple Polygons Christian Knauer, Günter Rote B 05-05 April 2005 Shortest Inspection-Path Queries in Simple Polygons Christian Knauer, Günter Rote Institut für Informatik,
More informationADVANCED DATA STRUCTURES FOR SURFACE STORAGE
1Department of Mathematics, Univerzitni Karel, Faculty 22, JANECKA1, 323 of 00, Applied Pilsen, Michal, Sciences, Czech KARA2 Republic University of West Bohemia, ADVANCED DATA STRUCTURES FOR SURFACE STORAGE
More informationA FRAMEWORK FOR REAL-TIME TERRAIN VISUALIZATION WITH ADAPTIVE SEMI- REGULAR MESHES
A FRAMEWORK FOR REAL-TIME TERRAIN VISUALIZATION WITH ADAPTIVE SEMI- REGULAR MESHES Lourena Rocha, Sérgio Pinheiro, Marcelo B. Vieira, Luiz Velho IMPA - Instituto Nacional de Matemática Pura e Aplicada
More informationCluster Analysis. Isabel M. Rodrigues. Lisboa, 2014. Instituto Superior Técnico
Instituto Superior Técnico Lisboa, 2014 Introduction: Cluster analysis What is? Finding groups of objects such that the objects in a group will be similar (or related) to one another and different from
More informationPolynomial Degree and Lower Bounds in Quantum Complexity: Collision and Element Distinctness with Small Range
THEORY OF COMPUTING, Volume 1 (2005), pp. 37 46 http://theoryofcomputing.org Polynomial Degree and Lower Bounds in Quantum Complexity: Collision and Element Distinctness with Small Range Andris Ambainis
More informationParallel Analysis and Visualization on Cray Compute Node Linux
Parallel Analysis and Visualization on Cray Compute Node Linux David Pugmire, Oak Ridge National Laboratory and Hank Childs, Lawrence Livermore National Laboratory and Sean Ahern, Oak Ridge National Laboratory
More informationHIGH ORDER WENO SCHEMES ON UNSTRUCTURED TETRAHEDRAL MESHES
European Conference on Computational Fluid Dynamics ECCOMAS CFD 26 P. Wesseling, E. Oñate and J. Périaux (Eds) c TU Delft, The Netherlands, 26 HIGH ORDER WENO SCHEMES ON UNSTRUCTURED TETRAHEDRAL MESHES
More informationA Tool for Generating Partition Schedules of Multiprocessor Systems
A Tool for Generating Partition Schedules of Multiprocessor Systems Hans-Joachim Goltz and Norbert Pieth Fraunhofer FIRST, Berlin, Germany {hans-joachim.goltz,nobert.pieth}@first.fraunhofer.de Abstract.
More informationOn Reliability of Dynamic Addressing Routing Protocols in Mobile Ad Hoc Networks
On Reliability of Dynamic Addressing Routing Protocols in Mobile Ad Hoc Networks Marcello Caleffi, Giancarlo Ferraiuolo, Luigi Paura Department of Electronic and Telecommunication Engineering (DIET) University
More informationEntropy-Based Collaborative Detection of DDoS Attacks on Community Networks
Entropy-Based Collaborative Detection of DDoS Attacks on Community Networks Krishnamoorthy.D 1, Dr.S.Thirunirai Senthil, Ph.D 2 1 PG student of M.Tech Computer Science and Engineering, PRIST University,
More informationProject Time Management
Project Time Management Study Notes PMI, PMP, CAPM, PMBOK, PM Network and the PMI Registered Education Provider logo are registered marks of the Project Management Institute, Inc. Points to Note Please
More informationChapter 6 Experiment Process
Chapter 6 Process ation is not simple; we have to prepare, conduct and analyze experiments properly. One of the main advantages of an experiment is the control of, for example, subjects, objects and instrumentation.
More informationPLAANN as a Classification Tool for Customer Intelligence in Banking
PLAANN as a Classification Tool for Customer Intelligence in Banking EUNITE World Competition in domain of Intelligent Technologies The Research Report Ireneusz Czarnowski and Piotr Jedrzejowicz Department
More informationAPP INVENTOR. Test Review
APP INVENTOR Test Review Main Concepts App Inventor Lists Creating Random Numbers Variables Searching and Sorting Data Linear Search Binary Search Selection Sort Quick Sort Abstraction Modulus Division
More information